to me it's pretty much a show stopper to not be on a recent version of Hive, 0.7 at least
I like having things like views, multi column count distincts and a HAVING clause . YMMV of course it's not at all hard to install hive on ec2 after using whirr to set the cluster up setting up the mysql metastore can be more painful though https://ccp.cloudera.com/display/CDHDOC/Hive+Installation I think if you are considering best practices one of the most important decisions you need to make is how aggressively you are going to upgrade. Guy On Tue, Aug 30, 2011 at 9:34 PM, jiang licht <licht_ji...@yahoo.com> wrote: > That's true. If its not acceptable, then one can replace hive in emr > cluster w/ latest version and then reuse it or build image from latest > hadoop and hive ... > > Michael > ------------------------------ > *From:* Igor Tatarinov <i...@decide.com> > *To:* user@hive.apache.org; jiang licht <licht_ji...@yahoo.com> > *Sent:* Tuesday, August 30, 2011 8:26 PM > *Subject:* Re: Hive in EC2 > > The only caveat is that you are at Amazon's mercy in terms of the latest > version of Hive. Also, they have their own versioning so EMR Hive's latest > version 0.7.1 could be Apache Hive's 0.6.5 - I am not even sure. Basically, > don't expect the latest Hive features to be available. > > igor > decide.com > > On Tue, Aug 30, 2011 at 6:25 PM, jiang licht <licht_ji...@yahoo.com>wrote: > > Recommend Amazon ElasticMapReduce. Otherwise, it costs you time to prepare > and set up hadoop and hive package for running on ec2. EMR does heavyweight > lifting work for you and still allow you option to customize your hadoop and > hive by pointing to their property files in xml (e.g. in S3). EMR also > allows your hive job to run in batch mode (through emr client command tools > or amazon consoler) or in interactive mode for test/debug purpose. Another > benefit of using EMR/hive is that its hive has enhanced features otherwise > not available, s.a., passing parameters from command line, loading > partitions automatically from S3 instead of loading them individually, etc. > Here's a link to emr faq and you may take a look at the answer to "Are there > new features in Hive specific to Amazon Elastic MapReduce?" > > http://aws.amazon.com/elasticmapreduce/faqs/ > > Michael > ------------------------------ > *From:* "Aggarwal, Vaibhav" <vagg...@amazon.com> > *To:* "d...@hive.apache.org" <d...@hive.apache.org>; "user@hive.apache.org" > <user@hive.apache.org> > *Sent:* Tuesday, August 30, 2011 11:51 AM > *Subject:* RE: Hive in EC2 > > You could also choose to look at Amazon ElasticMapReduce. > It allows you to provision an EC2 cluster of your choice preinstalled with > Hive and Hadoop. > > https://cwiki.apache.org/confluence/display/Hive/HiveAmazonElasticMapReduce > > Thanks > Vaibhav > > -----Original Message----- > From: MIS [mailto:misapa...@gmail.com] > Sent: Monday, August 29, 2011 11:03 PM > To: user@hive.apache.org; hive > Subject: Hive in EC2 > > Hi, > > Can somebody point me to production level setup of Hive in EC2. The intent > is to know the setup best practices being employed. > > Thanks. > > > > > >
